A chatbot on your website or WhatsApp can take a lot of routine work off your front desk. A badly scoped one can also frustrate patients, give wrong information and create more work than it saves. The difference is almost always in what you allow it to answer, and in how quickly it hands the conversation to a person.
Here’s the line we draw when we build chatbots for clinics and hospitals.
Questions a chatbot handles well
These are predictable, have one correct answer, and don’t depend on the patient’s health:
- Opening hours and holiday closures
- Location, parking and directions
- Which services, specialties or doctors you offer
- Which insurance plans or payment methods you accept (if this is a fixed list)
- How to prepare for common visits, using text your clinicians have approved, word for word
- Booking, rescheduling and cancelling, either directly or by collecting the request for staff
- Where to find forms, results portals or contact details
In most clinics these make up a large share of incoming messages. Count your own: export a week of WhatsApp or website messages and tag each one. That number tells you how much a chatbot can realistically save.
Questions it should never answer on its own
- Anything about symptoms, diagnosis or treatment. Even “is this normal?” is a clinical question.
- Medication and dosage questions.
- Test results or anything from a patient’s record.
- Prices that depend on an assessment (“how much will my treatment cost?”).
- Complaints, billing disputes or upset patients.
- Anything urgent. The bot should recognise words like chest pain, bleeding or emergency and immediately show your emergency instructions and number.
For all of these, the right answer is a fast, polite handover: “That’s a question for our team. I’ve passed your message to them and they’ll reply during opening hours,” plus what to do if it’s urgent.
Rules-based, AI, or both?
- Rules-based (menu buttons and fixed flows): predictable and easy to approve. Best for booking and simple FAQs.
- AI (language models): understands free-text questions phrased in many ways. Useful for FAQs, but it must be restricted to your approved information and told to hand over anything outside it.
- Both: the most practical setup. Buttons for booking, AI for understanding free-text FAQs, strict rules for handover and emergencies.
Whichever you choose, test it with your staff before patients see it, using real questions from your inbox.
Designing the handover
A good handover is what makes patients trust the bot:
- The patient can type “talk to a person” at any time.
- The full conversation goes to a shared team inbox, so nobody asks the patient to repeat themselves.
- Out of hours, the bot says honestly when someone will reply.
- Staff can take over a conversation, and the bot stays quiet while they do.
Patient data
Keep the chatbot’s knowledge to public information about your clinic. Collect only what a booking needs (name, contact, preferred time), limit who can read conversations, and check where your chatbot provider stores message data against your own regulatory obligations.
Measuring whether it works
Track a few simple numbers for the first month: how many conversations the bot finished on its own, how many it handed over, how many booking requests it collected, and what patients asked that it couldn’t answer. That last list tells you what to add next.
Thinking about a chatbot for your clinic or hospital?
We design and build WhatsApp and website chatbots for healthcare, with human handover built in, and connect them to your booking and reminder system. Talk to us about the questions your team answers most.